This project presents a multi-agent framework for simulating the Federal Open Market Committee (FOMC) decision-making process and predicting interest rate changes.

TherLiD is an open-source dataset developed to investigate racial and ethnic disparities in infrared thermometry within ICU settings.

A study on deepfake video detection, evaluating transfer learning, fine-tuning, and model generalization across two datasets, DFDC and FakeAVCeleb.

3rd Place Duke AI Hackathon: Alex and Mia is a real-time emotion-processing app designed to assist individuals with Alexithymia in recognizing emotions and improving social interactions.

Top 3 finish in predicting auto insurance claim costs using machine learning models, with a focus on the CatBoost algorithm and feature importance.